The training money arrived. The programmes didn't.

On 1 July the United States turned on Workforce Pell, extending federal grant aid to short-course workforce training for the first time. It was a decade in the making, community colleges lobbied hard for it, and the Department of Education and the Congressional Budget Office both expect it to reach 100,000 or more students by autumn 2027. Then colleges opened the rulebook and started measuring their own courses against it. At St Paul College in Minnesota, the certified nursing assistant class that trains people for the exact roles the region is short of runs to 112 instructional hours. The floor for eligibility is 150. Not one of the college's workforce programmes qualifies.

The box is specific. A programme must run between 150 and 599 clock hours, last at least eight weeks and fewer than fifteen, lead to a recognised credential and connect to a further academic pathway. It must then clear three outcome gates: a completion rate of at least 70 per cent, a job placement rate of at least 70 per cent within 180 days, and a positive return on investment, meaning median earnings of completers exceed tuition and fees plus 150 per cent of the federal poverty line. And it must pass through two doors, state approval by the governor working with the state workforce board, then federal approval by the Secretary of Education. Nothing in that list is unreasonable. Together they exclude a very large share of what community colleges currently teach, because most short-course training was designed around what an employer would pay for, not around a clock-hour band.

The approval machinery is also nowhere near ready. Of 52 jurisdictions tracked, 19 have an operational approval process as of early July, meaning an application window open, a pilot running or a list going to Washington. Twenty-six are somewhere in progress with a lead agency named or a working group meeting. Seven have no verifiable public action anyone has been able to find. The in-demand frameworks that colleges need before they can even apply had been published by only about a dozen states as of the end of June, and they do not resemble each other: Florida named 31 career certificate programmes, from phlebotomy to commercial vehicle driving; Michigan published 267 eligible occupations and left it to colleges to work out which programmes to build against them. The federal policy officer at the Association of Community College Trustees has told colleges to treat this year as a pilot, and her honest guess at when money actually reaches a student is January. Most students will be waiting until spring.

The interesting failure, though, is not the clock hours. Course length is a design problem and it is already being solved: St Paul is fusing its nursing assistant course with a medication aide certificate to clear the 150-hour floor, and North Idaho College, which happens to have built longer courses, is submitting five programmes including welding and HVAC without changing a thing. Curriculum can be re-cut in a semester. What cannot be re-cut in a semester is the evidence. To prove completion, 180-day placement and median earnings for a non-credit certificate, a college needs a data system that follows people who were never enrolled as degree students into jobs it never tracked. Many do not have one. Some are resorting to surveys, which means ringing former students and asking them where they work. A programme can be in a high-demand field, be exactly the right length, and still fail because nobody can evidence what happened to the people who finished it.

This is where the story stops being American. Every agency, everywhere, gets the same question from a serious investor: can you train 300 of these people, to this standard, in this timeframe? The honest answer is usually a brochure about a local college. The answer that closes a project is a named programme with a completion rate, a placement rate and a median starting wage attached to it. Workforce Pell has simply put a price on that distinction and made it visible. It says, in effect, that a training programme is only real to the extent it can prove its outcomes, and it has quietly reclassified outcomes tracking from an evaluation chore into the gate that determines whether money flows. Agencies in Ireland, Morocco, Poland and the Philippines have been building exactly this evidence layer for years, and it is why their workforce claims survive due diligence.

So the practical work for an economic development officer this quarter is not lobbying, and it is not waiting. It is an audit. Take the training programmes you cite in pitches, which for most agencies is a shortlist of a dozen or fewer, and find out three things about each: its clock hours, whether the state has put its occupation on the in-demand list, and whether the college can produce completion and placement numbers without conducting a phone survey. The first is a fixable design flaw. The second is a lobbying target with a named decision-maker. The third, if the answer is no, is the reason your workforce pitch will keep losing to a region whose answer is yes.